Previous research has suggested that while relaxation training may have little effect on an unselected sample of patients with non-insulin dependent diabetes mellitus (NIDDM), it may be effective in improving glucose tolerance in the subset of individuals who show high Trait Anxiety. The primary purpose of the proposed investigation is to determine if a relaxation- based stress management therapy (stress management) can significantly improve metabolic control in a group of individuals with NIDDM pre-selected for high Trait Anxiety. We will examine the acute and chronic effects of stress management on diabetes control in high and low Trait Anxiety patients with NIDDM, determine the relationship between changes in perceived anxiety and changes in blood glucose control in these patients and, in addition, we will perform exploratory analyses to determine if patients with repressive coping style (low Trait Anxiety and high social desirability), also show improvements in glucose metabolism with stress management training. Methods. A 2 X 2 design includes a high anxious group and a low anxious group--half of each group randomly assigned to receive the intervention of stress management training. The training will be conducted in 5, small-group classes at weekly intervals during the first 2 months of the study year. All subjects will attend 5 weekly classes--general diabetes education (for controls) and diabetes education plus stress management training (for experimentals). All subjects will be followed for 1 year, with assessments at baseline, 2 months, 4 months, 6 months, 9 months (phone) and 12 months. Variables include glucose tolerance, glucose stimulated insulin response, plasma glucagon levels, fasting plasma glucose/insulin, hemoglobin A1c, fructosamine, perceived anxiety/stress level, mean/variability of home-monitored blood glucose, diet analysis, learning [of diabetes education and stress management content], exercise level. In addition to these, history and physical exam, NEO Personality Inventory and Marlowe Crown Social Desirability Scale will be administered at baseline only. Men and women with NIDDM not requiring insulin treatment, will participate. Using norms for an adult general medical-surgical population, potential subjects will be classified into tertiles of high, low or moderate Trait Anxiety. Persons with scores in the high Trait Anxiety or low Trait Anxiety tertiles will be asked to complete the entire study. We estimate 360-400 people will need to complete the STAI in order to define 120 patients (each) in the upper and lower tertiles of Trait anxiety. Significance. This project will determine if a relaxation-based stress management program can have a clinically meaningful long-term impact on blood glucose control in high trait-anxious patients. If such an effect can be demonstrated, then we will be able to pre-select treatment responsive individuals. Success in applying this individual difference paradigm to the treatment of diabetes will suggest that similar approaches with other diseases, such as hypertension, should be tried. Plans. Because subjects scoring either >51 or <27 on the original criteria for Trait anxiety were few in number, we have enrolled all screened [interested and willing] subjects regardless of anxiety scores. We continued to overselect as possible from the subject pool in the high and low score ranges. To include a larger number of patients and generate more useful data, we considered Trait anxiety to be a continuous variable.